U.S. patent number 4,325,706 [Application Number 06/178,486] was granted by the patent office on 1982-04-20 for automated detection of platelets and reticulocytes in whole blood.
This patent grant is currently assigned to Ortho Diagnostic Systems Inc.. Invention is credited to Russell J. Gershman, W. Peter Hansen, Alan M. Hochberg, J. Garland O'Connell.
United States Patent |
4,325,706 |
Gershman , et al. |
April 20, 1982 |
Automated detection of platelets and reticulocytes in whole
blood
Abstract
A sample of whole blood is stained with an acridine orange
reagent, and is analyzed rapidly, a cell at a time, in a flow
cytometry system having a sample stream dimension in the range of
expected red cell dimensions. Red florescence and forward scatter
data is utilized first to discriminate a cell from noise, and then
to distinguish platelets from reticulocytes and red cells. The red
cell and reticulocyte data is subjected to a correction such as
rotational coordinate shift, and the shifted data are, by means of
statistical procedures, utilized to determine threshold criteria
separating red cells from reticulocytes, and to enumerate the cells
on that basis.
Inventors: |
Gershman; Russell J.
(Middleboro, MA), Hansen; W. Peter (Middleboro, MA),
Hochberg; Alan M. (Cambridge, MA), O'Connell; J. Garland
(Waltham, MA) |
Assignee: |
Ortho Diagnostic Systems Inc.
(Raritan, NJ)
|
Family
ID: |
22652721 |
Appl.
No.: |
06/178,486 |
Filed: |
August 15, 1980 |
Current U.S.
Class: |
435/6.16;
356/39 |
Current CPC
Class: |
G01N
15/1459 (20130101); G01N 33/5094 (20130101); G01N
2015/0084 (20130101); G01N 2015/0076 (20130101); G01N
2015/0073 (20130101) |
Current International
Class: |
G01N
15/14 (20060101); G01N 33/50 (20060101); G01N
001/30 (); G01N 021/25 (); G01N 033/50 () |
Field of
Search: |
;23/23B ;422/68,81,82
;356/39,40 ;424/3 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Serwin; Ronald
Attorney, Agent or Firm: Ciamporcero, Jr.; Audley A.
Claims
We claim:
1. A method of identifying and enumerating platelet and
reticulocyte cells in whole blood comprising the steps of:
(a) providing an aliquot from the blood to be studied;
(b) staining at least the reticulocyte and platelet cells of said
aliquot with an acridine orange reagent;
(c) passing at least a portion of said aliquot, substantially a
cell at a time through an area of focused optical stimulation, said
area having a cross-sectional dimension comparable to the expected
dimension of red cells;
(d) detecting light scattered by, and fluorescent light stimulated
from cells in said area;
(e) identifying platelet identification criteria as a linear
function of detected scattered light and detected fluorescent
light;
(f) discriminating platelets from reticulocytes and red blood cells
based on said identification criteria;
(g) providing an orthogonality correction at least as to
fluorescent light detected from individual cells, based on scatter
and fluorescent light detected from the cells;
(h) defining fluorescent light threshold criteria separating red
cells from reticulocytes; and
(i) identifying as reticulocytes those cells, platelets excluded,
detected as issuing fluorescent light greater than said threshold
criteria.
2. A method as described in claim 1 wherein said step of providing
an orthogonality correction comprises the steps of:
(a) assembling a scatter versus fluorescence distribution at least
for red cell singlets and multiplets;
(b) characterizing said red cells as a linear function of scatter
and fluorescence; and
(c) correcting the fluorescent light coordinate of cells in said
distribution by a correction proportional to said linear
function.
3. A method as described in claim 2 and further including
assembling reticulocyte cell data in said distribution, and wherein
said correcting step includes identical correction for said
reticulocyte cell data.
4. A method as described in claim 2 or claim 3 wherein said
characterizing step includes evaluating the angular disparity
between said linear function and the scatter axis of said
distribution, and wherein said correcting step includes providing a
radially constant, polar correction of fluorescence and scatter
coordinates of cells of said distribution, as a function of said
angular disparity.
5. A method as described in claim 1, and further including the step
of excluding platelet fluorescence and scatter data from said
correction step and steps subsequent thereto.
6. A method as described in claim 1 wherein said defining step
comprises the steps of:
(a) assembling a distribution of red and reticulocyte cells versus
associated fluorescent light stimulations;
(b) identifying the peak of said distribution;
(c) fitting a Gaussian curve to the portion of said distribution
below said peak; and
(d) detecting the fluorescent level at which said distribution,
above said peak, deviates from said Gaussian curve by a
predetermined statistical function, said detected level being the
basis for said threshold criteria.
7. A method as described in claim 1 wherein said threshold criteria
comprise adoption of said detected fluorescent level as a nominal
threshold between red cells and reticulocytes, and providing
predetermined correction to cell counts on at least one side of
said threshold as a function of said Gaussian curve and said
statistical function.
Description
FIELD OF THE INVENTION
This invention relates to the automated analysis of blood cells,
and more particularly to automated approaches for identifying and
enumerating platelets and reticulocytes.
BACKGROUND OF THE INVENTION
Detection and enumeration of reticulocyte cells has provided
extensive challenge to designers and manufacturers of automated
hematology instruments. Reticulocytes are precursors to red blood
cells, and hence the term reticulocyte embraces the evolution and
development of the cell whereby the red blood cell is generated.
Hence, a cell which is clearly a red cell precursor today will be a
red blood cell in a few days more, and will of course mature during
the interim. It is, therefore, quite difficult to define objective
criteria whereby reticulocytes may be effectively discriminated
from red cells. Correspondingly, criteria which have heretofore
been developed tend to involve quite subjective interpretations,
whereby even the most meticulous manual counts, for example
utilizing microscopic optical scanning techniques, will yield
results of moderate to extensive disparity.
It is an object of the present invention to provide an automated,
effective, highly repeatable approach to discrimination of
reticulocytes from red blood cells.
An approach to automated hematology which is increasingly finding
acceptance as the preferred approach is one often designated as
optical flow cytometry. Such systems employ a hydrodynamically
focused channel through which blood cells are passed extremely
rapidly, one at a time. The constriction of the channel is
illuminated by precisely focused light, for example coherent
radiation from a laser. Much can be determined by analysis of light
scattered by the cells, and if the blood sample has been treated
with specific staining agents, still more can be determined by
suitable analysis of fluorescent light stimulated from a stained
cell or other fluorescent material passing through the focusing
zone.
It is an object of the present invention to utilize the principles
of optical flow cytometry automatically and repeatably to identify
and enumerate platelets, reticulocytes, and red blood cells in a
whole blood sample.
SUMMARY OF THE INVENTION
In accordance with the principles of the present invention, a whole
blood aliquot, and typically quite a small one, is treated with
acridine orange reagent to stain at least the reticulocytes and
platelets, and the sample is passed through an optical flow
cytometry flow cell having a substantially narrowed hydrodynamic
focal region. Red fluorescent light and forward scattered light are
sensed, and based on threshold comparisons, occurrence of a cell in
the focal region is noted. A level shift allows characterization of
fluorescence related noise, whereupon platelets are discriminated
from reticulocytes and red cells based on identification criteria
which are a linear function of detected scattered light and
detected red fluorescent light (i.e. a straight line threshold in a
scatter versus red fluorescence histogram). Thereupon, an
orthogonality correction is conducted as to scatter data of
individual cells based on fluorescent light detected from the same
cell (i.e. coordinate rotation), whereupon a threshold level is
developed effectively to separate the red cells from
reticulocytes.
DESCRIPTION OF THE DRAWINGS
FIG. 1 shows a stylized version of a commercially available flow
cytometric apparatus, which may be adapted for utilization in
accordance with the principles of the present invention.
FIGS. 2, 3, and 4 show pump systems whereby sample fluid stream
dimensions in the flow cell are narrowed considerably, enabling
operation in accordance with the principles of the present
invention.
FIG. 5 shows in functional block diagrammatic form a system for
processing scattered light and fluorescence emissions in order to
discriminate and enumerate platelets and reticulocytes in
accordance with the principles of the present invention.
FIGS. 6A and 6B, and 7A and 7B, 8A and 8B and 9 show various
histograms illustrating sequences of operation for the embodiment
of FIG. 5 in accordance with the principles of the present
invention.
BEST MODE FOR CARRYING OUT THE INVENTION
Referring first to FIG. 1, there is shown a stylized functional and
structural representation of apparatus which may be utilized in
accordance with the principles of the present invention. In fact,
the apparatus of FIG. 1 depicts a particular system available
commercially under the trade designation CYTOFLUOROGRAPH.RTM.,
which is sold by Ortho Instruments, 410 University Avenue,
Westwood, Mass. 02090. The apparatus of FIG. 1 incorporates the
principles of flow cytometry for cell analysis, and includes
capacity for sensing fluorescent response of cells to specific
types of illumination.
Focal to the FIG. 1 apparatus is a flow channel 106, wherein cells
in liquid suspension are passed, in single file and at a rapid rate
(e.g. 2500 cells per second) through a sensing zone. The sensing
zone is defined by the intersection of cell flow and an incident
light beam, typically focused coherent light from a gas laser. As
the cell passes through the sensing zone, it interacts with
incident light in a variety of ways. Some light, of course, is
absorbed by the cell, other light is scattered at relatively narrow
angles to the axis of incident light, and still other light is
scattered at angles quite divergent from the axis of incident
light, for example at right angles to the incident light.
Furthermore, depending upon the nature of the cell itself, and any
dyeing or staining to which the cell may previously have been
subjected, fluorescence emissions may occur.
Accordingly, photosensors located at various orientations with
respect to the cell stream and the incident laser light permit
detection of a unique set of responses for each given type of cell.
Thus FIG. 1 includes an argon ion laser 101 and a helium neon laser
102, with the coherent light emitted by each being variously
deflected via mirrors 103 and 104 and a lens 105 to the sensing
zone of the flow channel 106. As is known in the art, the cell
sample stream is carried in laminar fashion within a flowing fluid
sheath, to insure that but a single cell will be illuminated in the
sensing zone at a given time. Hence, as each cell is illuminated by
light from the lens, interaction of the cell with the light may be
sensed.
As shown in FIG. 1, an extinction sensor 108 detects the amount of
light blocked by the cell, and forward light scatter is detected by
photosensors 109 and 110 approximately in a cone of half-angle
20.degree.. Electrical signals generated by the sensors 108, 109
and 110 are coupled to amplifiers 120 and 121, which present
electrical signals of suitable amplitude and the like for
subsequent analysis and/or display.
In the apparatus of FIG. 1, light which is emitted from the cell by
virtue of a fluorescence response is sensed at right angles both to
the direction of cell flow and to the axis of incident light. A
spherical mirror 125 and a condenser lens 107 collects this light
approximately in a cone of half-angle 20.degree., and couples this
light through an aperture 111, successively to a dichroic mirror
112 and to a second mirror 113. A first color filter 114 (e.g. to
pass relatively long wavelength light) conveys select light from
the dichroic mirror 112 to photosensor 117 (e.g. photomultiplier
tube). A second filter 115 selectively passes light of a different
color (e.g. relatively short wavelength light) from the second
mirror 113 to a second photosensor 116. Electrical signals from
sensors 116 and 117, in the form of pulses corresponding to light
from respective cells, are coupled to amplifiers 118 and 119,
thereby also to produce signals which are adapted for suitable
processing.
As shown in the FIG. 1 embodiment, a sensor selector 122 generates
output histograms utilizing signals from the amplifiers 118 through
121. For example, one useful form of output is a plot of amplitude
of red fluorescence, from sensor 117, against amplitude of green
fluorescence, from sensor 116. Such a histogram is shown at display
123, with each point on the histogram representing an individual
cell. Clusters or aggregates of points on the histogram represent
groups of cells of similar type. Quite evidently, those of ordinary
skill in the art find it useful variously to generate histograms of
narrow forward angle scatter versus intensity of green
fluorescence, narrow forward angle scatter versus axial light
extinction, and so forth.
In accordance with the principles of the present invention, it is
highly desirable severely to constrict the cross-section of the
sample fluid stream at the focal point of the flow channel 106, for
example from the more common 20 microns or so in cross-section,
down essentially to the cell size of 6 to 7 microns or so. Such
narrow fluid stream vastly reduces the noise emergent from the
irradiated sample, and hence renders considerably more practical
the proposition of making the measurements and discriminations
called for in accordance with the principles of the present
invention. That is, it will be appreciated that noise as a function
of spurious matter passing through the focal point of the fluid
stream occurs roughly in proportion to the radius of the stream.
Such is also the case with respect to fluorescence related noise,
due at least in part to free stain in the stream. Hence the
desirability of reducing the fluid stream dimension as much as
possible, while still accomodating the cells under examination.
Conventionally, flow cytometry systems have employed a pressure
system for delivering sheath fluid and sample fluid into the flow
channel. The stream narrowing criteria established for the present
invention is easily achieved utilizing a precision sample pump for
delivery of sample fluid, and a differential pressure regulator and
associated restricter for regulation of the flow of pressure driven
sheath fluid. A preferred configuration, which also is the subject
matter of concurrently filed application U.S.S.N. 178,489 of J. G.
O'Connell entitled "Controlled Hydrodynamic Flow in Flow Cytometry
Systems", is shown in FIGS. 2, 3 and 4.
Referring to FIG. 2, there is shown symbolically a flow cell
assembly 106 wherein a fluid sheath 201 surrounds a sample fluid
flow 202. Such constructions are well-known to those of ordinary
skill in the art. Sheath flow is provided from a sheath fluid
supply bottle 203, which is pressurized from an air pressure supply
204 via an air pressure regulator 205. Sheath fluid from the bottle
203 passes to a differential pressure regulator 206 and a
temperature controlled restriction 207 which the regulator 206
controls. A constant pressure drop therefore is maintained across
the restricter 207, irrespective of pressure changes either up or
down sheath flow stream. The result is a constant mass flow rate of
sheath fluid 201 through the cell 106.
Sample is supplied through a valve 209, and is pressurized by a
precision gas tight syringe 210 whose plunger 312 displaces the
sample fluid volume. The plunger 312 is stroked by a precision lead
screw 211, driven by a permanent magnet synchronous gear motor 213
in conjunction with a universal coupling 212. The sample fluid 202
enters the flow cell assembly 106 from a fluid damper 208.
The development of flow within the cell is obtained by first
supplying pressure to the sheath fluid supply bottle 203. Once the
mass flow rate of the sheath fluid is established, sample fluid is
injected into the system through the sample valve 209. The sample
fluid is, in turn, metered through the damper 208 and into the flow
cell 106 as the syringe plunger 210 is displaced by operation of
the motor 213 and 211. The result is a sample fluid stream 202
encased by sheath fluid 201 in the flow cell 106. The linear
velocity of the sample fluid is determined solely by the sheath
fluid mass flow rate, while the mass flow rate of the sample fluid
is determined solely by the displacement of the sample pump. The
sample fluid stream, within the flow cell, has dimensions
determined by both the velocity and mass flow rate of the sample
fluid. The system thus produces a sample fluid stream of very small
dimension which is uniform and constant over time.
FIG. 3 shows a cross-sectional view of a preferred configuration
for the motor 213, coupling 212, screw 211 and syringe 210. Support
members 305, 306, and 314 are mounted to a base 315, and thereby
establish relative spatial relationships between the respective
parts. A synchronous motor 213 is affixed to the leftmost upright
support 305, and the motor drive shaft 307 meets a universal
coupling 212, which in turn drives the precision lead screw 211
between brackets 306 and 314. A rail 317 is carried between support
members 306 and 314 intermediate the lead screw 211 and syringe
210, and a lead screw nut 311 is connected to a trolley 316, which
rides on rail 317 and, on the opposite side thereof is connected to
syringe plunger rod 312. The plunger 312 is matable within the
syringe barrel 313.
It will therefore be seen that rotational motive force from
synchronous motor 213 is transferred by coupling 212 to the lead
screw 211. As lead screw 211 is turned, the lead screw nut is
translated thereupon, and correspondingly the syringe plunger 312
is moved into or out of the syringe barrel 313.
FIG. 4 shows an alternative approach to the system set forth in
FIG. 2, wherein the sheath fluid line includes a pump means
substantially similar to that used in the sample fluid line. Hence,
in FIG. 4, the fluid line does not include the differential
pressure regulator 206 and restrictor 207, instead employing a
motor 413, lead screw 411, and syringe combination 410/415 in
similar fashion to the apparatus included in the sample line.
Essentially, the FIG. 4 embodiment, like the FIG. 2 embodiment,
provides sheath fluid at a constant mass flow rate. The embodiment
of FIG. 4, however, by employing an instantly activated motor-pump
rather than a thermally responsive pressure regulator and
restrictor, as a more rapid start up time. That is, the thermal
aspects of the pressure regulator 206 and restrictor 207 of the
FIG. 2 embodiment requires several minutes initially to heat up to
achieve operational level. The embodiment of FIG. 4 has no such
limitation.
Inasmuch as the normal operation of flow cytometry systems requires
larger volumes of sheath fluid than sample fluid, the FIG. 4
embodiment no doubt will employ a syringe 410 having a larger
volume capacity than the one 210 utilized by the sample fluid line.
Likewise, the FIG. 4 embodiment employs separate sheath fluid lines
and sample fluid lines which employ respective operating parameters
adapted to the desired constant mass flow rate of the respective
fluids to be delivered to the fluid cell assembly 106; however,
components of the respective lines are functionally analagous to
one another in each respect.
In partial summary, then, utilization of the pump and pressure
system of FIG. 2 or 4, in conjunction with a system such as shown
symbolically in FIG. 1, facilitates processing of whole blood
samples in accordance with the principles of the present invention
to discriminate and enumerate platelets and reticulocytes.
An important aspect of the principles of the present invention is
to provide a suitable stain for platelets and reticulocytes whereby
the coupling of focused coherent light to the stained cells will
result in emissions of fluorescent light at predictable, controlled
wavelengths in order to complete processing in conjunction with
scatter measurements, and thereby to discriminate platelets and
reticulocytes from one another and from red cells. One such dye is
acridine orange.
In a concurrently filed copending application of Peter J. Natale,
U.S. Ser. No. 178,481, entitled "Reagent for automated counting of
platelets and/or reticulocytes", assigned to the assignee hereof,
there is described a preferred reagent for utilization in
accordance with the principles of the present invention. The
copending Natale application utilizes the dye known as acridine
orange, and employs a formulation which takes advantage of the
cationic/anionic reaction between acridine orange and
ribo-deoxyribo-nucleic compounds. These reactions result in
complexes which fluorescent in the red or green wavelength range,
when excited with appropriate radiation. Since both platelets and
reticulocytes contain forms of ribonucleic compounds, they can be
distinguished from other cells and each other by their respective
light scatter and fluorescent characteristics, in accordance with
the principles of the present invention.
The reagent described by Natale, in order to render the activity of
the acridine orange compound relatively optimal, includes
paraformaldehyde for stabilization of the cytoplasmic membranes so
that natural cell sizes and shapes are maintained, for enhancement
of acridine orange entry to and within cell matrices, and for
enhancement of the stripping of ribonucleic proteins from the
respective nucleic acid for the exposure of additional reactive
sites. Additionally, the Natale reagent includes a citrate
constituent for chelation of calcium, thus preventing platelet
clumping or aggregation, for cationic chelation, thus enhanceing
cellular cationic exchange resulting in increased acridine orange
entry, for buffering effects near isoelectric pH of interfering
proteins, and for maintainance of isotonicity for preservation of
cell shape and size.
Thus, for preparation of blood in accordance with the principles of
the present invention, and aliquot of whole blood is taken, and at
least a portion of the aliquot is provided with adequate amounts of
acridine orange stain, preferably in a reagent complex of the type
disclosed in the aforementioned Natale application, but
alternatively other suitable dye compositions. Thus, as the blood
sample, having appropriately stained reticulocyte and platelet
cells, is coupled to apparatus of the type shown in either FIGS. 2
or 4, and thence to a flow cytometry system of the sort shown in
FIG. 1, cell scatter measurements and fluorescence measurements are
taken. In a preferred embodiment red florescence is the
fluorescence parameter of greatest significance. Processing in
accordance with the principles of the present invention may be
appreciated upon consideration of the block diagrammatic system
setforth in FIG. 5, and the histograms setforth in FIGS. 6A and 6B,
7A and 7B, 8A and 8B, and 9.
Referring, then, to FIG. 5, the red fluorescence signals, for
example from sensor 117 and amplifier 119 of FIG. 1, and the
forward scatter signals from sensors 109 and 110 and amplifier 120
of FIG. 1, are coupled to a cell detection network 501. It is the
purpose of the cell detection network to distinguish presence of
the cell in the flow channel 106 from spurious signal which might
otherwise be mistaken as a cell. The presence of a cell is deemed
to occur when a linear combination of the input parameters, red
fluorescence at line 502 and forward scatter at line 503, exceeds a
threshold level. Hence, the cell detection network 501 simply
includes separate red fluorescence and forward scatter thresholds,
the achievement of both of which can occur substantially only by
presence of a cell in the illumination zone of the flow channel
106. Since the detection network 501 serves chiefly to eliminate
noise, the red fluorescence and forward scatter thresholds will
essentially be empirical values, conditioned on the precise design
parameters associated with the flow channel being employed, the
nature of the dyes being utilized to stain the cells, and the
amount of free or uncombined dye passing through the flow channels
and thereby subject to the issuance of fluorescent noise.
Additionally, the presence of spurious matter in the sheath or
sample fluids may contribute somewhat to noise. In practice,
establishment of these noise thresholds is neither a difficult nor
elaborate procedure, in accordance with the knowledge and ability
of those ordinary skill in the art.
When a cell is detected, noise threshold network 501 couples
enabling signals to separate pulse integration networks 504 and
505, which, as noted, perform a straightforward integration
process, and produce at their outputs an integrated signal
representing the area under respective red fluorescence and forward
scatter pulses delivered at inputs 502 and 503. The integrated
forward scatter signal is coupled directly to an analog to digital
converter 508, and also to still another threshold network 507,
which as noted conducts a decision process to distinguish certain
cell types. The red fluorescence integrated signal is provided with
a baseline offset at amplifier 506, and then is coupled both to the
analog to digital converter 508 and to the decision network 507.
The addition of the baseline offset at amplifier 506 may be
appreciated upon consideration of the histograms of FIGS. 6A and
6B. Both histograms represent a plot or distribution of scatter (on
the ordinate) versus red fluorescence (on the abscissa). No such
histogram is assembled at this point of the operation, but the
histograms, as shown, do illustrate the variety of data
combinations which may occur. Groupings corresponding to platelets,
reticulocytes, red cell doublets, and red cell singlets will be
seen, but it will be noted that the red cell singlet distribution
is clustered about the ordinate axis, and indeed the presence of
the red fluorescence noise, when combined on a cumulative basis,
may result in certain red cells having a negative or zero reading
unless an offset is added. Accordingly, after integration occurs at
504, the integrated red fluorescence signal is provided at
amplifier 506 with a fixed offset which in essence allows for
acquisition of the entire distribution of red fluorescent integrals
both above and below the baseline integral value. In particular,
the offset red fluorescence distribution of red cells and
reticulocytes allows accurate statistical procedures to be
followed, as discussed hereinafter. The effect of the offset may
readily be appreciated upon consideration of FIG. 6B.
A first cell decision process occurs at threshold network 507, in
essence distinguishing platelets from either red cells or
reticulocytes. In other words, the integrated scatter and red
fluorescence signal (which if actually accumulated point 507 would
appear as the clusters set forth in FIG. 6B), have a rather
distinct aggregation of platelets, well spaced from the remainder
of cells. Hence, a linear discriminant (i.e. a linear combination
of scatter and red fluorescence factors), will facilitate
identification of each forward scatter-red fluorescence combination
arriving to decision network 507, as either corresponding to
occurrence of a platelet in the flow channel 106, or occurrence of
another cell, which may either be a red cell or a reticulocyte, but
which in any event is not susceptible to determination at this
stage of the procedure.
Again, from consideration from FIG. 6B, it will be noted that there
is indeed a considerable disparity from the aggregation of points
corresponding to platelets, and those corresponding to the
remainder of the cells; hence, identification of a suitable linear
discriminant, for any individual system, will be a rather routine
matter, given the knowledge that such a discriminant is to be
drawn. If a cell's integrated red fluorescence-scatter coordinate
places it below the discriminant line, then that combination
indicates that the cell is a platelet and its red fluorescence
integral is coupled by a line 515 to a memory network 516, which
accumulates a cell count versus red fluorescence data in a manner
exemplified by the histogram set forth in FIG. 7A. It will be
appreciated that the histogram set forth in 7A is constituted by
defining a number of levels or channels along the red fluorescence
axis, and for each occurrence of a signal on line 515 having a red
fluorescence integral within such range, there is an incrementing
of the stored cell count for the associated level or channel.
Hence, FIG. 7A sets forth a symbolic presentation, in smooth or
analog form, of the information stored in memory 516.
In the event that a forward scatter-red fluorescence combination is
apprehended by threshold network 507 to be above the linear
discriminant shown symbolically in FIG. 6B, indicating that a
platelet cell has not been detected (but lacking the ability to
determine whether a reticulocyte or red cell has been detected),
the integrated red fluorescence pulse is coupled via line 514 to
yet another memory 519, which operates as described
hereinafter.
Signal digitization occurs at analog to digital converter 508,
producing 8 bit digitally encoded representations, on respective
output lines 512 and 513, of integrated forward scatter pulses, and
offset, integrated red fluorescence pulses.
An appropriate mode of A to D conversion is as follows. When a cell
is detected at 501 there is presented at 502 and 503 pulses, which
when integrated are each represented as a level equivalent to the
area under the associated pulse. By conducting a linear discharge
of the integration signal down to zero, a pulse may be formed, with
the width of the pulse proportional to the amplitude of the
integral. This pulse is then compared with a very rapid (e.g. 40
megahertz) signal, and the duration of the pulse is represented
digitally as the number of 40 megahertz cycles which occur during
the length of the total pulse. This number is presented at lines
512 and 513 as 8 bit words, i.e. to a granularity of 256
levels.
It is to be noted that the accuracy of the platelet count will
depend largely on the purposes for which that count is taken. For
example, in a "platelet only" mode, it may be necessary, or
desirable, to allow the user substantial discretion in order to
have the final platelet count meet the accuracy desired. For
purposes of reticulocyte count, however, a much less accurate
platelet count is necessary, it being desirable primarily to use
the platelet enumeration for purposes for discrimination of
reticulocytes, as described hereinafter.
For subsequent assembly of a red-cell/reticulocyte histogram, a
rotation correction is to be made based on a preliminary run of
signals. In order to establish this first distribution of red
fluorescence versus forward scatter, the precision of the digitized
red fluorescence and forward scatter signal optionally may be
reduced (e.g. from 8 bit to 6 bit). Thereupon, these 6 bit words
are passed into a memory 509 (e.g. 64 by 64) which allows for the
storage of information corresponding to a red fluorescence versus
forward scatter histogram. This data will be utilized as discussed
for purposes of evaluating the rotation correction to be had. FIG.
8A, depicts such a histogram of scatter versus red fluorescence,
and amply illustrates the need for a coordinate rotation or the
like correction process. The upper grouping on the histogram,
representing red cell multiplets such as doublets, along with the
singlet red cell distribution, is characterized by a positive slope
on the scatter versus red fluorescence distribution. Coordinate
rotation is therefore desirable for at least two reasons. First,
there is need to have the multiplet distribution vertically aligned
with the singlet distribution to facilitate curve fitting processes
(to be described hereinafter). Secondly there is need to have a
single threshold which will amply descriminate between
reticulocytes and all red cells, including multiplets.
The 64 by 64 distribution of red fluorescence versus scatter stored
in memory 509 may be thought of as having discrete "channels", each
memory location representing the intersection of a scatter channel
and a red fluorescence channel. A coordinate rotation is best
accomplished by investigating the red blood cell distribution
through each scatter channel in which it occurs. Generally, it is
known that red blood cells will occur within a given scatter range,
and this facilitates the investigation considerably. The basic
approach is, for each scatter channel in this range, to detect the
fluorescence channel having the peak of the distribution, and then,
by progressively extending the inquiry to adjacent fluorescence
channels, to evaluate the mean red fluorescence for that given
scatter channel. Such inquiry is conducted for each of the scatter
channels which carry the red blood cells distribution. The
aggregate of these separate means are processed, by simply curve
fitting procedures, to yield a single straight line (i.e. of the
form y=mx+b where x is red fluorescence, b is the y intercept, m is
the slope, and y is scatter) which represents the major axis of the
red cell distribution. Hence, with respect to a straight vertical
axis, the straight line fit also indicates the amount of coordinate
rotation which will be conducted. Simple principles of trigonometry
likewise enable the evaluation of an angle which characterizes the
slope m of the straight line.
It will be appreciated that the entire foregoing procedure,
designated only by block 570, could well be embodied by
specifically designed, hard wired apparatus including large numbers
of registers, encoders, multiplexers, and the like standard digital
processing equipment. In fact, however, it is far preferable to
conduct these procedures by means of suitably programmed digital
computers. Numerous such apparatus are commercially available and
in general use; one of particular attraction for use in accordance
with the principles of the present invention is available from Data
General Inc. under the trade name "Micronova", and features a 32K
memory, and thereby the facility to process 16 bit words. Having
described preferred rotation correction methods in detail in the
foregoing, it will be appreciated that software designers of
ordinary skill in the art may, depending upon the operating system
being employed, their own desires and facility in terms of
utilizing machine language and aspects of the commercial operating
system, accomplish the requisite tasks expeditiously. So also could
digital hardware engineers ply their skills to fabricate hardwired
versions of determination block 510.
Referring back to FIG. 5, then, the foregoing procedures yield an
angle through which the reticulocyte, and red cell (singlet and
multiplet) distributions are to be rotated. This angle is coupled
to the next block, designated "rotation process". In fact, given a
rotation angle (or linear correction function), the actual
correction is a relatively simple subtraction process.
Meanwhile, once the rotation angle has been evaluated at 510, the
data in memory 509, which was utilized for evaluation of the angle,
is less relevant. That is, while it may or may not be desirable
further to utilize such data, the essential evaluations to be made
in accordance with the principles of the present invention relate
to a conduct of the rotation process with respect to digitized red
fluorescent signals, and the accumulation of a one dimensional
(i.e. cell count versus red fluorescence) distribution of red blood
cells and reticulocytes.
It may, then, be desirable merely to clear the memory 509 once the
correction criteria have been developed. As shown in FIG. 5, the
digitized 8 bit word is subjected at 511 to a rotation correction,
and thence is passed to a memory 519 which assembles data
representing a cell count versus red fluorescence histogram of the
combined red blood cells and reticulocyte counts.
On the subject of the rotation itself, it will be noted that a
rotation as such, on a scatter versus red fluorescence
distribution, entails a correction with respect to both axes. That
is, in a true polar rotation, there will occur both an x and y
correction, although in accordance with the principles of the
present invention, the y correction may indeed be small. Hence, in
FIG. 5, the process is broadly characterized, it being understood
that the degree of attention to the x and y amplitude shifts, will
vary in accordance with the needs of the designer, and the speed of
the available processing apparatus. Indeed, some may wish not to
perform strict rotation, instead performing only an approximated
shift in parallel to the desired major axis. Such options are well
within the purview of the principles of the present invention.
Once the RBC/retic count versus red fluorescence distribution is
assembled in memory 519, it is appropriate to separate the red
blood cell count from the reticulocyte count. FIGS. 7B and 9
represent the data stored in Memory 519.
The principle source of noise in a system such as utilized in
accordance with the principles of the present invention, and
generally characterized in FIG. 1, is the sample stream itself.
That is, as discussed hereinbefore, extreme noise may arise from
spurious matter, uncombined fluorescent dyes or materials, and the
like in the stream. Essentially, this "white" or Gaussian noise,
and accordingly the data in Memory 519, in part mimics a Gaussian
curve, including the actual contribution of the irradiated red
cells, if any, and the noise generated attendant to fluorescence
generated from the stream.
Hence a preferred rationale for discriminating red blood cells from
reticulocytes is to fit a Gaussian curve to the lower of leftmost
portions of the FIG. 9 distribution, and once a proper Gaussian
curve has been fit to that data, to conduct statistical inquiries,
progressively rightwardly along the red fluorescence channels, to
determine the point at which the actual cell count curve begins to
diverge substantially (i.e. within given statistical criteria) from
the Gaussian curve. Such point is deemed a nominal threshold
between all red blood cells and reticulocytes. Above this
threshold, then, one may extrapolate the Gaussian curve to account
for the balance of red cells; excluding these, all other cells
above the threshold are safely deemed reticulocytes.
In greater detail, the process is performed as follows, reference
being made to FIG. 9 for purposes of clarification. In fact, FIG. 9
is a symbolic depiction of data stored in Memory 519 of FIG. 5.
First, the "mode" or fluorescence channel having peak red cell
count is identified. Thereupon, count data in all channels to the
left of the mode may be utilized to develop a Gaussian curve which
best fits the "half distribution", which curve will be employed for
the later half of the distribution. Thereupon, for each channel
beyond the mode and in sequence, the developed curve is extended to
find the count which would result were the cell counts in the
channels to follow a precise Gaussian distribution. For each such
channel, it is preferred to conduct statistical test to determine
the degree to which the actual count for the channel diverts from
the Gaussian calculation. Such fitting and comparison steps occur
sequentially, red fluorescence channel by red floresence channel,
above the mode of the distribution. A threshold is deemed to be
found whenever a comparison yields a difference from the expected
or Gaussian point, which exceeds the predetermined statistical
standards.
Once the threshold has been established, the count in each channel
therebeyond will have a red cell component indicated by the
extrapolated Gaussian curve, and a reticulocyte component, based on
the difference between that extrapolated curve and the actual count
in the channel. In fact, viewing FIG. 9, it will be noted that the
Gaussian contribution (i.e. red blood cells) quickly vanishes above
the threshold.
It will be appreciated that the threshold level, being selected as
a function of statistical criteria, will no doubt be recognized as
a point slightly after some reticulocytes have been embodied in the
counts. It may, accordingly, be desirable to "back off" by a number
of channels, determined by the designer, in order to pick up a few
reticulocytes which, prior to statistical identification of the
threshold, were dismissed as statistical aberration.
In FIG. 5, the foregoing operations are embodied in block 520,
which like block 510, is embodied as a suitably programmed computer
system, or as a specially designed hardwired mechanism.
It will be appreciated that the literature includes numerous curve
fitting approaches and techniques whereby a wide variety of curves,
including Gaussian functions, may be fit to given collections of
data. Indeed, software packages are commercially available for such
curve fitting operations. The principles of the present invention
are not directed per se to such a purely mathematical formula, and
are not intended to appropriate such purely mathematical
processing. Rather, it is contemplated in accordance with the
principles of the present invention that these mathematical
formulae, which are taught throughout the literature, may be
employed advantageously (so long as they may be accommodated by the
memory, speed, and the like parameters of the hardware system being
employed) in order to discriminate the types of cells being studied
in accordance with the principles of the present invention.
Likewise, numerous statistical tests are available when conducting
the channel by channel Gaussian fit versus actual count
comparisons. One such test which has been found suitable is the Chi
square test (one degree of freedom), but it will be appreciated
that numerous others may also be appropriate.
Therefore, total reticulocyte and red cell counts are evaluated at
520, and the respective counts coupled for a final coincidence
correction based upon the platelet counts.
The significance of the platelet correction may be appreciated by
considering FIG. 4A, together with the following precepts. First,
viewing FIG. 4A, it will be noted that the reticulocyte aggregation
likely has a scatter parameter which is in the same range as that
for a red blood cell, but is distinguished therefrom based on
divergent red fluorescence ranges. The reticulocytes also, however,
have a red fluorescence value roughly equivalent to that of
platelets, and are distinguished therefrom based largely upon
divergent scatter ranges. It is to be noted, however, that
platelets and red cells could possibly reside in the sample stream
so closely to one another that a red cell forward scatter signal
and a platelet red fluorescence signal might, from the standpoint
of system dynamics, be issued at substantially the same time.
Hence, there would be seen by the system a forward scatter-red
fluorescence combination which would indicate a reticulocyte in the
channel, rather than a coincident red cell-platelet event. The
platelet coincidence correction avoids, on a statistical basis,
error attendant to this phenomenon. Essentially, the platelet
coincidence correction at 521 is conducted by reducing the
reticulocyte count developed at 520 by a factor equivalent to the
likelihood of a platelet-red cell event occurring.
Perhaps exemplary numbers will illustrate this point. Assuming a
cell sample having raw (i.e. uncorrected) counts of n red cells,
0.01 n reticulocytes, and 0.05 n platelets, with the system
operating at such a rate that at any given time there is a
probability p that any cell is present. In such a situation, there
is substantially 0.05 p likelihood of a coincidence of platelet and
a red cell. In such event, the platelet correction at 521 would be
accomplished by reducing the reticulocyte count by 0.05 p.
Generally, for systems reasonably within the contemplation of the
principles of the present invention, the factor "p" may be in the
range of of 2 to 3 percent. Hence the platelet coincidence
correction may not be required for certain applications.
It will be appreciated that the foregoing has set forth preferred
and illustrative embodiments of the principles of the present
invention, but that numerous alternative embodiments may occur to
those of ordinary skill in the art without departure from the
spirit or scope of the present invention.
* * * * *